1. Frye v. United States (1923) — The “General Acceptance” Rule

Core Issue

Whether scientific evidence is admissible in court.

Facts

A defendant tried to introduce a systolic blood pressure deception test (early polygraph-like method).

Court’s Holding

The court rejected it because the method was not generally accepted in the scientific community.

Legal Principle

Scientific evidence is admissible only if it has “general acceptance” in its field.

Relevance to Epidemiology

• Early epidemiological studies were often excluded if they were “novel”
• Courts heavily relied on consensus rather than data strength
• This created barriers for statistical and population-based science

Importance

This rule dominated U.S. courts for decades before being partially replaced.

2. Daubert v. Merrell Dow Pharmaceuticals (1993) — Modern Scientific Evidence Standard

Core Issue

Whether expert epidemiological testimony linking Bendectin (anti-nausea drug) to birth defects was admissible.

Facts

Plaintiffs claimed Bendectin caused limb defects based on scientific studies.

Supreme Court Holding

The Court rejected the strict Frye standard and introduced a new test.

Daubert Standard (Key Factors)

Judges must evaluate:

• Testability of the theory
• Peer review and publication
• Known error rate
• General acceptance (still relevant but not controlling)

Relevance to Big Data Epidemiology

This case opened the door for:

• Large epidemiological datasets
• Meta-analyses
• Statistical causation models
• Regression-based risk studies

Importance

This is the foundation of modern scientific evidence law in the U.S.

3. General Electric Co. v. Joiner (1997) — Judge as “Gatekeeper”

Core Issue

How much discretion judges have in accepting or rejecting scientific evidence.

Facts

Plaintiff alleged exposure to PCB chemicals caused lung cancer. Experts used animal studies and limited epidemiological inference.

Supreme Court Holding

Courts can exclude expert testimony if there is too large a gap between data and conclusion.

Key Principle

Judges act as “gatekeepers” of scientific reliability.

Relevance to Epidemiology

• Courts may reject big data studies if methodology is weak
• Correlation alone is not enough for causation
• Data must be logically connected to the conclusion

Importance

Strengthened judicial control over complex statistical evidence.

4. Kumho Tire Co. v. Carmichael (1999) — Daubert Applies to All Experts

Core Issue

Whether Daubert applies only to scientific experts or also technical/experience-based experts.

Facts

A tire blowout case where an expert used visual inspection rather than statistical analysis.

Supreme Court Holding

Daubert applies to all expert testimony, including:

• Engineers
• Medical experts
• Statistical analysts
• Epidemiologists

Relevance to Big Data Epidemiology

This case ensured:

• Epidemiological experts must meet reliability standards
• Courts can scrutinize statistical methods (sampling, bias, confounding factors)
• Big data alone is not automatically valid

Importance

Expanded scientific scrutiny beyond pure science to applied fields.

5. Agent Orange Litigation (In re “Agent Orange” Product Liability Litigation, 1980s)

Core Issue

Whether exposure to Agent Orange caused cancer and other diseases in veterans.

Evidence Used

• Military exposure records
• Large epidemiological studies
• Veteran health databases
• Statistical comparisons of disease rates

Court Outcome

The court rejected most claims due to insufficient causal proof despite large datasets.

Key Reasoning

• Epidemiological association did not prove causation
• Confounding environmental factors existed
• Data was statistically inconclusive

Relevance to Big Data Epidemiology

This case is a classic example of:

• Massive datasets not being enough without causal linkage
• Difficulty of proving low-probability health risks in courts

6. Bendectin Litigation (Multiple U.S. Cases, 1980s–1990s)

Core Issue

Whether Bendectin caused birth defects.

Evidence Used

• Multiple epidemiological studies involving thousands of pregnancies
• Meta-analyses
• Case-control studies

Court Outcome

Courts consistently ruled:

• No reliable causal connection proven

Significance

This litigation strongly influenced the Daubert decision.

Relevance to Big Data Epidemiology

• One of the earliest large-scale uses of epidemiological databases in court
• Showed courts prefer replicated statistical evidence over isolated studies

7. Asbestos Litigation (Ongoing Mass Tort Cases)

Core Issue

Whether asbestos exposure causes mesothelioma and lung diseases.

Evidence Used

• Occupational health databases
• Longitudinal epidemiological studies
• Exposure modeling over decades
• Industry-wide disease registries

Court Findings

Unlike Bendectin:

• Causation was widely accepted
• Epidemiological consensus strongly supported liability

Key Legal Principle

When:

• Exposure is well-documented AND
• Disease incidence is significantly elevated in exposed populations
  → Courts accept epidemiological causation strongly.

Relevance to Big Data Epidemiology

This is one of the strongest examples where:

• Big data + consistent epidemiology = legal causation established

Overall Legal Principles Emerging from These Cases

Across all these cases, courts have developed consistent rules:

1. Correlation is not causation

Large datasets showing association are not enough.

2. Methodology matters more than size

Even “big data” can be rejected if:

• Sampling is biased
• Confounders are not controlled
• Statistical models are weak

3. Judges act as scientific gatekeepers

Under Daubert/Joiner/Kumho, courts actively evaluate:

• Statistical validity
• Epidemiological methods
• Logical reasoning

4. Consensus strengthens admissibility

If multiple independent epidemiological studies agree, courts are more likely to accept them.

5. Causation requires legal + scientific convergence

Courts require:

• Epidemiological evidence
• Biological plausibility
• Exposure linkage
• Consistency across studies

Conclusion

Big data epidemiology has transformed courtrooms from narrative-based reasoning to data-driven scientific evaluation. However, courts do not treat statistical scale as truth by itself. Instead, they demand methodological rigor, reproducibility, and logical causation.